论文部分内容阅读
在复杂网格环境下,MapReduce调度任务如何有效地利用共享可用资源实现最短的时间内完成分配任务,这样的任务分配问题是一个NP难题.提出一种混合差分粒子群算法(HDE-PSO)求解任务调度问题.新的混合差分-粒子群算法(HDEPSO)引入了DE算法的突变和交叉算子,克服传统PSO算法容易陷入局部最优解的缺陷,以增加算法的全局寻优能力.通过实验证明该HDE-PSO算法比传统PSO和DE算法具有更好的收敛性和寻优能力,并能更充分的利用共享资源.
In a complex grid environment, how to effectively share the available resources and achieve the task of allocation in the shortest time by MapReduce scheduling task is an NP problem.This paper proposes a Hybrid Differential Particle Swarm Optimization (HDE-PSO) algorithm Task scheduling problem.The new hybrid difference-particle swarm optimization (HDEPSO) introduces the mutation and crossover operator of DE algorithm to overcome the flaws of the traditional PSO algorithm that is easy to fall into the local optimal solution to increase the global optimization ability of the algorithm.Through experiments It is proved that the HDE-PSO algorithm has better convergence and optimization ability than traditional PSO and DE algorithms and can make full use of shared resources.